Author Topic: Bitcoin Lightning Network  (Read 3026 times)

0 Members and 1 Guest are viewing this topic.

Offline mike623317

  • Hero Member
  • *****
  • Posts: 637
    • View Profile
I need to try and digest these slides.

Thanks for giving us your thoughts BM.

Offline Thom

Thanks Troglodactyl for reminding me of http://bytemaster.bitshares.org/article/2015/01/20/The-Minimal-Requirement-for-Decentalization/, which I will read again. The 3 of the 4 links you provided are comparisons, focusing on BitShares superior merits relative to other approaches, not so much on the data we have from our own (the first) DPoS experiment. I'm looking for an in depth review and analysis of DPoS 1.0 in practice. It's been months since I've looked at the blog articles, so I may have forgotten that such an analysis has been done and discussed already. I don't think so, but if that's the case I do apologize right here, right now.

Remember that in the next update it will not be fixed to 101 witnesses.  Shareholders can buy as much decentralization as they think is beneficial and worth the cost.

EDIT: +5% DSN...
With the witnesses being payed by transaction fees ... it doesn't really matter how many there are from the companies perspective .. does it?

This point has been raised, and I think we should discuss what the correct number of witnesses should be and how we make that determination. If witnesses will be paid solely from transaction fees we could begin to look at the tradeoff between transaction fees and witness pay. The variable factors that go into that are:
  • VPS server costs
  • Income from transaction fees
  • How decentralized (by physical location, legal jurisdiction and witness operator)

Items 1 & 3 are closely related. How should we establish criteria for decentralization? For example, how many delegates should be allowed in a given area, or could be run by the same person?

Nobody should be required to work for free, so what is a reasonable pay rate for labor, and how to we accommodate the variability in the cost of VPS in the various locations we deem are important to maintain?

We could take the highest cost VPS servers in use now and use that as a basis for a fixed cost, knowing all other instances would be less than that, or we could devise some type of variable reimbursement scheme so witnesses get 100% of blocks produced until the server costs are paid then drop down to the fixed pay rate all witnesses will have (for their time to maintain the nodes). I think it would be better to just have a fixed pay rate that includes both VPS costs and the node maintenance labor costs.

What should the fee to register as a witness be? Should that be reimbursed or not?

If we took ONLY the current 3% delegates right now and ranked them on reliability, if they publish feeds, where they are located and what legal jurisdiction they are bound to, will there be a sufficient number of them to adequately support the minimum level of decentralization shareholders want? If so the fee to be set isn't terribly important in the near term, but needs to be established so that if someone wants to campaign for a witness to compete with those in place currently they can.

What about support for seed nodes and chain servers, should they be treated like a standard witness? Could / should standard block producing nodes be combined with chain / seed servers, and if so should they be allowed to retain more pay to compensate for higher RAM / DISK costs?

There are a lot of questions that need to be explored in how best to establish these things. What I'm saying in part is why wait until 6/8 to start to discuss them?

As to relying on the shareholders to know "how much decentralization" would be best, yes they obviously need a voice in that but I think their voice should be supported by sound technical facts, as well as legal considerations.  There's no reason I can think of why we must wait until Monday to start thinking about that.

Injustice anywhere is a threat to justice everywhere - MLK |  Verbaltech2 Witness Reports: https://bitsharestalk.org/index.php/topic,23902.0.html

Offline xeroc

  • Board Moderator
  • Hero Member
  • *****
  • Posts: 12922
  • ChainSquad GmbH
    • View Profile
    • ChainSquad GmbH
  • BitShares: xeroc
  • GitHub: xeroc
Remember that in the next update it will not be fixed to 101 witnesses.  Shareholders can buy as much decentralization as they think is beneficial and worth the cost.

EDIT: +5% DSN...
With the witnesses being payed by transaction fees ... it doesn't really matter how many there are from the companies perspective .. does it?

Yes, it does.  The transaction rate and total fees are distinct from number of witnesses, so increasing the number of witnesses reduces the pay of each and thus reduces incentive to compete for each position.

Determining number of witnesses thus allows shareholders to establish a balance between decentralization and network performance.  Once the TPS rate hits the limit of common personal computers, this will be more of an issue.  We're nowhere near this issue now, but the devs are building in long term scalability.
Yes ... but the DAC as a company does not need to pay extra for higher decentralization (i.e. more witnesses) .. that's what I wanted to say ..

Sure .. the more pieces of the pie are getting smaller for increasing number of witnesses and thus the incentives ..

Offline Troglodactyl

  • Hero Member
  • *****
  • Posts: 960
    • View Profile
Remember that in the next update it will not be fixed to 101 witnesses.  Shareholders can buy as much decentralization as they think is beneficial and worth the cost.

EDIT: +5% DSN...
With the witnesses being payed by transaction fees ... it doesn't really matter how many there are from the companies perspective .. does it?

Yes, it does.  The transaction rate and total fees are distinct from number of witnesses, so increasing the number of witnesses reduces the pay of each and thus reduces incentive to compete for each position.

Determining number of witnesses thus allows shareholders to establish a balance between decentralization and network performance.  Once the TPS rate hits the limit of common personal computers, this will be more of an issue.  We're nowhere near this issue now, but the devs are building in long term scalability.

Offline xeroc

  • Board Moderator
  • Hero Member
  • *****
  • Posts: 12922
  • ChainSquad GmbH
    • View Profile
    • ChainSquad GmbH
  • BitShares: xeroc
  • GitHub: xeroc
Remember that in the next update it will not be fixed to 101 witnesses.  Shareholders can buy as much decentralization as they think is beneficial and worth the cost.

EDIT: +5% DSN...
With the witnesses being payed by transaction fees ... it doesn't really matter how many there are from the companies perspective .. does it?

Offline Troglodactyl

  • Hero Member
  • *****
  • Posts: 960
    • View Profile
Remember that in the next update it will not be fixed to 101 witnesses.  Shareholders can buy as much decentralization as they think is beneficial and worth the cost.

EDIT: +5% DSN...

Offline BunkerChainLabs-DataSecurityNode


https://lightning.network/

At the end of the day, the Lightning Network will end up just as centralized as any other network due to pure economics. 

Conclusion:  a complex solution compared to BitShares approach that ultimately results in similar centralization and fees and is only fit for transfers.

I read this and my takeaways are:
  • Larimer believes all networks become centralized due to economics
  • The comparison between bitcoin's lightning and BitShares leaves BitShares in a less favorable light than ever before
  • Why hasn't Larimer discussed this "all networks become centralized" economic theory before, such as in his blog?
  • Doesn't point #1 ultimately lead us back to a broken, corruptible and centralized clone of the existing mainstream financial system? If not what stops that from happening?
  • Does Larimer no longer believe centralization of a global financial ecosystem is a bad thing?
  • What are the implications of bitcoin lightning on the future of BitShares? Does it impact the roadmap of 2015 or 2016?
  • Following up on delulo's question, how can you have privacy on a public ledger? WHAT is to be held as private and still maintain the value of a public ledger?

Quote
At the end of the day, the Lightning Network will end up just as centralized as any other network due to pure economics.

Conclusion:  a complex solution compared to BitShares approach that ultimately results in similar centralization and fees and is only fit

I interpret this as Dan saying that the Lightning Network will end up centralized and is only fit for transactions - as compared to the BitShares approach which is not.
The blog posts linked by Troglodactyl explain why BitShares has 101 witnesses, and accepts that some degree of centralization is inevitable, but that 101 witnesses is sufficient decentralization to keep the network secure from corruption. Acknowledging these factors now allows BitShares to direct funding elsewhere rather than lose money attempting to achieve full decentralization where every user participates in network validation.
The end user cannot be expected to maintain bandwidth and processing power to index and relay 100k's of transactions per second.

101 is dated.. the witnesses capacity will be unlimited and determined by the voting blockchain. It could get bigger, or smaller.
+-+-+-+-+-+-+-+-+-+-+
www.Peerplays.com | Decentralized Gaming Built with Graphene - Now with BookiePro and Sweeps!
+-+-+-+-+-+-+-+-+-+-+

Offline Permie

  • Hero Member
  • *****
  • Posts: 606
  • BitShares is the mycelium of the financial-earth
    • View Profile
  • BitShares: krimduss

https://lightning.network/

At the end of the day, the Lightning Network will end up just as centralized as any other network due to pure economics. 

Conclusion:  a complex solution compared to BitShares approach that ultimately results in similar centralization and fees and is only fit for transfers.

I read this and my takeaways are:
  • Larimer believes all networks become centralized due to economics
  • The comparison between bitcoin's lightning and BitShares leaves BitShares in a less favorable light than ever before
  • Why hasn't Larimer discussed this "all networks become centralized" economic theory before, such as in his blog?
  • Doesn't point #1 ultimately lead us back to a broken, corruptible and centralized clone of the existing mainstream financial system? If not what stops that from happening?
  • Does Larimer no longer believe centralization of a global financial ecosystem is a bad thing?
  • What are the implications of bitcoin lightning on the future of BitShares? Does it impact the roadmap of 2015 or 2016?
  • Following up on delulo's question, how can you have privacy on a public ledger? WHAT is to be held as private and still maintain the value of a public ledger?

Quote
At the end of the day, the Lightning Network will end up just as centralized as any other network due to pure economics.

Conclusion:  a complex solution compared to BitShares approach that ultimately results in similar centralization and fees and is only fit

I interpret this as Dan saying that the Lightning Network will end up centralized and is only fit for transactions - as compared to the BitShares approach which is not.
The blog posts linked by Troglodactyl explain why BitShares has 101 witnesses, and accepts that some degree of centralization is inevitable, but that 101 witnesses is sufficient decentralization to keep the network secure from corruption. Acknowledging these factors now allows BitShares to direct funding elsewhere rather than lose money attempting to achieve full decentralization where every user participates in network validation.
The end user cannot be expected to maintain bandwidth and processing power to index and relay 100k's of transactions per second.
JonnyBitcoin votes for liquidity and simplicity. Make him your proxy?
BTSDEX.COM

Offline Troglodactyl

  • Hero Member
  • *****
  • Posts: 960
    • View Profile

https://lightning.network/

At the end of the day, the Lightning Network will end up just as centralized as any other network due to pure economics. 

Conclusion:  a complex solution compared to BitShares approach that ultimately results in similar centralization and fees and is only fit for transfers.

I read this and my takeaways are:
  • Larimer believes all networks become centralized due to economics
  • The comparison between bitcoin's lightning and BitShares leaves BitShares in a less favorable light than ever before
  • Why hasn't Larimer discussed this "all networks become centralized" economic theory before, such as in his blog?
  • Doesn't point #1 ultimately lead us back to a broken, corruptible and centralized clone of the existing mainstream financial system? If not what stops that from happening?
  • Does Larimer no longer believe centralization of a global financial ecosystem is a bad thing?
  • What are the implications of bitcoin lightning on the future of BitShares? Does it impact the roadmap of 2015 or 2016?
  • Following up on delulo's question, how can you have privacy on a public ledger? WHAT is to be held as private and still maintain the value of a public ledger?

http://bytemaster.bitshares.org/article/2015/01/12/Decentralization-Scalability-and-Fault-Tolerance-of-Bitcoin/
http://bytemaster.bitshares.org/article/2015/01/09/How-to-Measure-the-Decentralization-of-Bitcoin/
http://bytemaster.bitshares.org/article/2015/01/20/The-Minimal-Requirement-for-Decentalization/
http://bytemaster.bitshares.org/article/2015/01/13/Decentralization-of-Nxt-vs-BitShares/

You mean like all these blog posts?  I think these issues have been discussed pretty thoroughly, and resolving them is almost the whole point of DPOS.
« Last Edit: June 06, 2015, 03:55:51 pm by Troglodactyl »

Offline Thom


https://lightning.network/

At the end of the day, the Lightning Network will end up just as centralized as any other network due to pure economics. 

Conclusion:  a complex solution compared to BitShares approach that ultimately results in similar centralization and fees and is only fit for transfers.

I read this and my takeaways are:
  • Larimer believes all networks become centralized due to economics
  • The comparison between bitcoin's lightning and BitShares leaves BitShares in a less favorable light than ever before
  • Why hasn't Larimer discussed this "all networks become centralized" economic theory before, such as in his blog?
  • Doesn't point #1 ultimately lead us back to a broken, corruptible and centralized clone of the existing mainstream financial system? If not what stops that from happening?
  • Does Larimer no longer believe centralization of a global financial ecosystem is a bad thing?
  • What are the implications of bitcoin lightning on the future of BitShares? Does it impact the roadmap of 2015 or 2016?
  • Following up on delulo's question, how can you have privacy on a public ledger? WHAT is to be held as private and still maintain the value of a public ledger?
« Last Edit: June 06, 2015, 03:17:49 pm by Thom »
Injustice anywhere is a threat to justice everywhere - MLK |  Verbaltech2 Witness Reports: https://bitsharestalk.org/index.php/topic,23902.0.html

Offline santaclause102

  • Hero Member
  • *****
  • Posts: 2486
    • View Profile
Since we are claiming BitShares can scale to 100,000 transactions per second I thought it would be fair to weigh in on what the Bitcoin core devs plan for scaling:

https://lightning.network/

I watched the video presentation of the slides and found the entire thing fascinating.  Here are my takeaways:

1. It functions a lot like a "collateralized" ripple network.  To make a payment you must find a route between the sender and receiver.
2. It requires "interactivity" and active participation of all links in the route, including the receiver.
3. It would reintroduce privacy to the blockchain.
3. While the simple example of  A->B->C->D appears to be simple enough where "B" and "C" are well known hubs.  If you replace A and D with 1000A's and 1000D's then the combinatoric complexity of the "un-broadcast transaction" between B and C can get rather large, especially when you consider all of the possible intermediate states.  B and C would probably end up setting up many parallel payment channels to prevent any one channel from becoming too complex. 
4. Even on the "lightning network" the time it would take to negotiate all of the handshakes and routing would be a second or more assuming nodes were distributed around the world. 

Their definition of scalability is:   7 billion people making 2 transactions per day each.   162,000 transactions per second.  50 Mbyte/sec "best case"

BitShares can handle that on a single computer at 20 Mbyte/sec which is well within reach of many data centers.  Internet2 can handle 12 Gigabyte per second bandwidth.

So here is my overall conclusion:

1) The Lightning Network would make an excellent currency if the complexity can be properly managed because it would be both fast and private.
2) To send a payment to an "offline" receiver would require hitting the blockchain. 
3) It will not work for markets on a blockchain
4) it would be "buggy" because long chains can have many failure points that end up "locking up funds" for longer than intended and would require some kind of reputation system due to the cost of opening a channel with a bad node. 
5) Economics of scale would drive most transactions through a small set of hubs because they can offer lower fees and have high reliability (no hangs waiting for a timeout).  These nodes would end up being like a 101 delegates and require much more than 50Mbyte/sec due to all of the handshaking required. 
6) While the number of transactions that end up hitting the blockchain will be fewer, the size of the transactions hitting the blockchain will be much larger.

At the end of the day, the Lightning Network will end up just as centralized as any other network due to pure economics. 

Conclusion:  a complex solution compared to BitShares approach that ultimately results in similar centralization and fees and is only fit for transfers.
Thanks for sharing it.

How much privacy would this introduce if measured with the definition of privacy (false privacy is no privacy etc.) you gave in the last (or second last) mumble hangout?

Tuck Fheman

  • Guest
3. It would reintroduce privacy to the blockchain.

 +5%

BitShares can handle that on a single computer at 20 Mbyte/sec which is well within reach of many data centers.

 +5% +5%

Conclusion:  a complex solution compared to BitShares approach that ultimately results in similar centralization and fees and is only fit for transfers.

BitShares ftw!

Offline bytemaster

Since we are claiming BitShares can scale to 100,000 transactions per second I thought it would be fair to weigh in on what the Bitcoin core devs plan for scaling:

https://lightning.network/

I watched the video presentation of the slides and found the entire thing fascinating.  Here are my takeaways:

1. It functions a lot like a "collateralized" ripple network.  To make a payment you must find a route between the sender and receiver.
2. It requires "interactivity" and active participation of all links in the route, including the receiver.
3. It would reintroduce privacy to the blockchain.
3. While the simple example of  A->B->C->D appears to be simple enough where "B" and "C" are well known hubs.  If you replace A and D with 1000A's and 1000D's then the combinatoric complexity of the "un-broadcast transaction" between B and C can get rather large, especially when you consider all of the possible intermediate states.  B and C would probably end up setting up many parallel payment channels to prevent any one channel from becoming too complex. 
4. Even on the "lightning network" the time it would take to negotiate all of the handshakes and routing would be a second or more assuming nodes were distributed around the world. 

Their definition of scalability is:   7 billion people making 2 transactions per day each.   162,000 transactions per second.  50 Mbyte/sec "best case"

BitShares can handle that on a single computer at 20 Mbyte/sec which is well within reach of many data centers.  Internet2 can handle 12 Gigabyte per second bandwidth.

So here is my overall conclusion:

1) The Lightning Network would make an excellent currency if the complexity can be properly managed because it would be both fast and private.
2) To send a payment to an "offline" receiver would require hitting the blockchain. 
3) It will not work for markets on a blockchain
4) it would be "buggy" because long chains can have many failure points that end up "locking up funds" for longer than intended and would require some kind of reputation system due to the cost of opening a channel with a bad node. 
5) Economics of scale would drive most transactions through a small set of hubs because they can offer lower fees and have high reliability (no hangs waiting for a timeout).  These nodes would end up being like a 101 delegates and require much more than 50Mbyte/sec due to all of the handshaking required. 
6) While the number of transactions that end up hitting the blockchain will be fewer, the size of the transactions hitting the blockchain will be much larger.

At the end of the day, the Lightning Network will end up just as centralized as any other network due to pure economics. 

Conclusion:  a complex solution compared to BitShares approach that ultimately results in similar centralization and fees and is only fit for transfers.

For the latest updates checkout my blog: http://bytemaster.bitshares.org
Anything said on these forums does not constitute an intent to create a legal obligation or contract between myself and anyone else.   These are merely my opinions and I reserve the right to change them at any time.